Overview
I am supervising projects for BSc, MMath and MSc students in the area of statistics and data science. These allow you to gather experience in independent learning and research, scientific programming and data analysis and, most importantly, in scientific writing.
If you are an MMath student in year 3 you already have a mandatory double-term thesis project scheduled for year 4 and therefore you should not take a year 3 thesis project.For MMath and MSc students these projects are mandatory, however for BSc students they are optional. Below you find some information to decide whether a year 3 project under my supervision is the right choice for you, and also how to apply.
Should I do a BSc project?
By their nature thesis projects are demanding and challenging on many different levels. Consequently, successfully completing a student project constitutes a high level of personal achievement and will greatly add to your skills learned as part of your degree.
Commonly, for most BSc students taking a regular taught module in year 3 is a better and more straightforward choice than a thesis project. However, if you are considering to go into postgraduate research (PhD) then you should definitely do a year 3 project and write a thesis. The reason is that when you later apply for a PhD you will need to be able to demonstrate that you have prior project and research experience and also skills in scientific writing. Furthermore, by doing a year 3 project you will be able to find out whether this direction is the right path for you.
What are typical topics for projects under your supervisions?
The general topic for my student projects is "exploring the interface between statistics and machine learning". Under this umbrella in the past specific topics have been deep learning, reinforcement learning, manifold learning and nonlinear dimension reduction, etc. For an detailed overview see the list of student thesis projects supervised by me.
In all cases the key idea is to investigate statistical foundations of modern machine learning methods. As a methods-driven project you will learn new methodology - new as in outside the normal curriculum for a BSc student but typically also new in terms of current research. You will study it from original literature, present it in your own words, evaluate its performance in computer simulations, analyse relevant data sets, interpret the results and also compare with more traditional methods.
Note that my projects generally are not about analysing a particular data set. However, all academics in the Statistics group offer BSc thesis projects, and some of them do have projects that are of the data analysis type that you may prefer.
How did previous Bsc students do?
All of the previous students did indeed very well in this project. All of them continued to postgraduate study on MSc and/or PhD level.
What are the conditions and requirements?
The main conditions are:
- You need to take the module "Multivariate Statistics and Machine Learning" (MATH38161) in Semester 1.
Desirable skills for the project are:
- experience in programming with data, in R and/or Python, possibly also in Julia and Stan,
- knowledge in Latex / Markdown to prepare scientific documents,
- strong motivation to learn and investigate new statistical methods and
- being self-driven with good time management skills.
How do I apply?
Please send me an email to inform me that you would like to do a BSc project with me. Subsequently, register your interest with Dr. Nikesh Solanki. He will then allocate students to supervisors, and inform supervisors and students.
I supervise around BSc/MMath students each year but there is usually much higher demand.
How does the project work?
Once the allocation has taken place I will meet with you at early in Semester 1. During the first meeting we will explore together your interests in statistical learning and determine the exact topic of the thesis project, tailored to your particular interests.
The project itself will then commence and we will have regular biweekly meetings. You will need to reserve one full working day per week to work on the thesis project. Typically, the first third is taken up by reading and studying, the second for data analysis and simulations, and the third for completing the writeup of the thesis. The submission of the thesis is at the end of Semester 2, with an oral presentation taking place during the exam period.